09-02-2013, 11:10 AM
Interference Mitigation in Heterogeneous cellular networks of macro and femto cells
[attachment=50305]
Abstract
We investigate the impact of co-channel deployment
of femtocells on the existing macro-cellular systems considering
the techniques to resolve the laud neighbor problem. There are
several approaches to this aim, e.g., femtocell power control,
interference coordination, opening access to femtocells and so on.
Among them, the coordinated scheduling, including power
control, and their impact will be the main focus of this w ork. In
the context of 3GPP-LTE, we examine under various operational
scenarios the performance in terms of the average and 5% worst
user throughput, a useful measure of fairness among users, both
for femto and macro cells. Although recent studies have shown
that co-channel femtocell has minor impact on the macrocell
performance in average sense, non-negligible percentage of users
may lose their opportunity to get satisfactory data service and,
hence, our focus will be put more on 5% worst users.
INTRODUCTION
Femtocell is a low transmission power base station (BS) with
direct high-speed internet connection. Its primary purpose is to
cover small isolated areas with insufficient macrocell coverage
(coverage hole) providing an access point to home or small
office users who can benefit from high signal to interference
ratio due to the proximity of the femto BS [1, 2].
Generally, it has been considered more feasible to introduce
femtocells on a different frequency resource from macrocell
due to coexistence problem of co-channel deployment in
which femtocell acts as an interference source specifically to
those who cannot access femtocells. One solution to avoid
such laud neighbor problem would be the self-power control,
i.e., femtocell first measures the signal power of the nearest
macro BS and adjusts its transmission power to a suitable
level. Since the shadowing can be highly correlated within a
small femtocell coverage area, the laud neighbor problem can
be resolved at least partially. Nevertheless, this approach does
not guarantee no macro user suffers from serious interferences.
Another solution can be an adaptive access policy as proposed
in [6], where femtocell is adaptively open to nearby macro
users under certain condition. This approach, however, may
raise another problem, e.g., security and/or privacy.
SYSTEM DESCRIPTION AND MODELING
The system to be considered consists of a multiple of
macrocells (macro B S) a nd many f emto c ells ( femto BS)
reside therein, all of which are assumed to be wire-connected
via backhaul to exchange information for instantaneous
scheduling decision, i.e., coordinated scheduling for multiple,
simultaneous OFDM based downlink transmissions. The
coordinated scheduling is performed by the central resource
manager which is assumed to be located at the center macro
BS and dedicated to coordinate the mutual interference among
the macrocell and femtocells. To simplify the problem, we
will assume single antenna transmission and that the backhaul
has enough bandwidth and negligible delay so that every BS
can instantaneously share their scheduling information and
channel information fed back from the MS associated to it. We
also assume that perfect channel information is available at
MS and is fed back without errors. We also assume the
centralized scheduling at the central resource manager (CRM)
located at the macro BS, i.e., every femto BS is “associated”
to the macro BS and the scheduling in macrocell and every
femtocell is arranged by the CRM.
Clustering issue
A cluster is defined as the set of BSs participating in the
multi-cell coordination and determining a cluster is an
important issue in designing the multi-cell coordination policy.
Two options can be considered for the clustering of BSs, either
(network-determined) static clustering or (user-determined)
dynamic clustering. In the former, the network determines the
cluster based on the geometrical network topology while, in the
latter, each user determines the cluster based on the link
strength measurement from nearby BSs. The static clustering is
plausible option in multi-cell coordination only between macro
cells since the network topology and the link strength are static.
In dynamic clustering, a cluster can be determined, for each
user, by selecting a multiple BSs whose received signal
strength is above some threshold relative to the signal strength
of its serving cell. In the cellular systems currently available in
the market, such measurement can be easily done along the cell
search procedure. Hence, in the multi-cell coordination scheme
to be discussed shortly, we will consider the user-determined,
dynamic clustering for coordination between macro and
femtocells.
Impact of Power Control
Now, we consider the throughput performance under femto BS
power control. First, in Fig 3, the throughput performances
under the RSRP based self power control were shown for Pboost
= 90, 80 and 70 dB. As depicted in this figure, one can observe
that the average and 5% worst user throughput in macro cell is
improved with smaller Pboost at the expense of the performance
in femto-cell. This is what we expected, i.e., by reducing the
power of femto BS, macro MSs get higher SINR for better
service quality. But, it is obtained only at the expense of the
throughput reduction in femto-cell. It is a tradeoff problem and
the value of Pboost must be set for minimum macro MS
throughput while maximizing the femto MS throughput.
CONCLUDING REMARKS
By deploying femto BS, noticeable interferences are induced
to the macro MSs and we observed from the simulation
results that some reduction in average macro cell throughput
and considerable reduction in 5% worst user throughput is
unavoidable without coordination. To effectively manage the
interferences between macro and femto-cell (and also
between femto-cells), coordinated scheduler must be
incorporated with a certain level threshold value. Simulation
results shows that the coordinated scheduling can bring
considerable 5% worst user throughput gain, especially for
macro MSs by reporting only a few (or several) dominant
interferer, which can be done aperiodically to the serving BS.
[attachment=50305]
Abstract
We investigate the impact of co-channel deployment
of femtocells on the existing macro-cellular systems considering
the techniques to resolve the laud neighbor problem. There are
several approaches to this aim, e.g., femtocell power control,
interference coordination, opening access to femtocells and so on.
Among them, the coordinated scheduling, including power
control, and their impact will be the main focus of this w ork. In
the context of 3GPP-LTE, we examine under various operational
scenarios the performance in terms of the average and 5% worst
user throughput, a useful measure of fairness among users, both
for femto and macro cells. Although recent studies have shown
that co-channel femtocell has minor impact on the macrocell
performance in average sense, non-negligible percentage of users
may lose their opportunity to get satisfactory data service and,
hence, our focus will be put more on 5% worst users.
INTRODUCTION
Femtocell is a low transmission power base station (BS) with
direct high-speed internet connection. Its primary purpose is to
cover small isolated areas with insufficient macrocell coverage
(coverage hole) providing an access point to home or small
office users who can benefit from high signal to interference
ratio due to the proximity of the femto BS [1, 2].
Generally, it has been considered more feasible to introduce
femtocells on a different frequency resource from macrocell
due to coexistence problem of co-channel deployment in
which femtocell acts as an interference source specifically to
those who cannot access femtocells. One solution to avoid
such laud neighbor problem would be the self-power control,
i.e., femtocell first measures the signal power of the nearest
macro BS and adjusts its transmission power to a suitable
level. Since the shadowing can be highly correlated within a
small femtocell coverage area, the laud neighbor problem can
be resolved at least partially. Nevertheless, this approach does
not guarantee no macro user suffers from serious interferences.
Another solution can be an adaptive access policy as proposed
in [6], where femtocell is adaptively open to nearby macro
users under certain condition. This approach, however, may
raise another problem, e.g., security and/or privacy.
SYSTEM DESCRIPTION AND MODELING
The system to be considered consists of a multiple of
macrocells (macro B S) a nd many f emto c ells ( femto BS)
reside therein, all of which are assumed to be wire-connected
via backhaul to exchange information for instantaneous
scheduling decision, i.e., coordinated scheduling for multiple,
simultaneous OFDM based downlink transmissions. The
coordinated scheduling is performed by the central resource
manager which is assumed to be located at the center macro
BS and dedicated to coordinate the mutual interference among
the macrocell and femtocells. To simplify the problem, we
will assume single antenna transmission and that the backhaul
has enough bandwidth and negligible delay so that every BS
can instantaneously share their scheduling information and
channel information fed back from the MS associated to it. We
also assume that perfect channel information is available at
MS and is fed back without errors. We also assume the
centralized scheduling at the central resource manager (CRM)
located at the macro BS, i.e., every femto BS is “associated”
to the macro BS and the scheduling in macrocell and every
femtocell is arranged by the CRM.
Clustering issue
A cluster is defined as the set of BSs participating in the
multi-cell coordination and determining a cluster is an
important issue in designing the multi-cell coordination policy.
Two options can be considered for the clustering of BSs, either
(network-determined) static clustering or (user-determined)
dynamic clustering. In the former, the network determines the
cluster based on the geometrical network topology while, in the
latter, each user determines the cluster based on the link
strength measurement from nearby BSs. The static clustering is
plausible option in multi-cell coordination only between macro
cells since the network topology and the link strength are static.
In dynamic clustering, a cluster can be determined, for each
user, by selecting a multiple BSs whose received signal
strength is above some threshold relative to the signal strength
of its serving cell. In the cellular systems currently available in
the market, such measurement can be easily done along the cell
search procedure. Hence, in the multi-cell coordination scheme
to be discussed shortly, we will consider the user-determined,
dynamic clustering for coordination between macro and
femtocells.
Impact of Power Control
Now, we consider the throughput performance under femto BS
power control. First, in Fig 3, the throughput performances
under the RSRP based self power control were shown for Pboost
= 90, 80 and 70 dB. As depicted in this figure, one can observe
that the average and 5% worst user throughput in macro cell is
improved with smaller Pboost at the expense of the performance
in femto-cell. This is what we expected, i.e., by reducing the
power of femto BS, macro MSs get higher SINR for better
service quality. But, it is obtained only at the expense of the
throughput reduction in femto-cell. It is a tradeoff problem and
the value of Pboost must be set for minimum macro MS
throughput while maximizing the femto MS throughput.
CONCLUDING REMARKS
By deploying femto BS, noticeable interferences are induced
to the macro MSs and we observed from the simulation
results that some reduction in average macro cell throughput
and considerable reduction in 5% worst user throughput is
unavoidable without coordination. To effectively manage the
interferences between macro and femto-cell (and also
between femto-cells), coordinated scheduler must be
incorporated with a certain level threshold value. Simulation
results shows that the coordinated scheduling can bring
considerable 5% worst user throughput gain, especially for
macro MSs by reporting only a few (or several) dominant
interferer, which can be done aperiodically to the serving BS.